Self-adaptive LED lamp control method and control system

Abstract

The invention discloses a self-adaptive LED lamp control method and system, and relates to the technical field of intelligent illumination. An LED lamp control platform is constructed, action feature recognition and face feature recognition are performed on video stream data acquired by an LED lamp terminal, action information of a target user is acquired, and the target user is marked as a registered user or a tourist user; when the target user is a tourist user, according to the multi-source heterogeneous data and action information stored in the current control period and the blockchain node, obtaining control parameters of an LED lamp terminal in an area where the tourist user is located; a user instruction prediction model is constructed, and when the target user is a registered user, control parameters of the LED lamp terminal in the area where the registered user is located are acquired according to the multi-source heterogeneous data and the action information of the current control period and the user instruction prediction model; the comfort level of the user is improved, the lighting parameters can be dynamically adjusted according to preferences of different users, and therefore the purposes of energy conservation and emission reduction are achieved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent lighting technology, and in particular to an adaptive LED lamp control method and control system. Background Art

[0002] Prior art CN107734746B "A control system for LED lamps", the control system of the LED lamps includes a lamp coding module, a chip coding module, a lamp query module, and a lamp control module, each of the LED lamps is assigned a different lamp address code, each of the different color types of LED chips in the LED lamps is assigned a different chip address code, and the same color type of LED chips in the LED lamps are assigned the same chip address code, the query instruction is composed of a lamp address code and a query command code, and the control instruction is composed of a lamp address code, a chip address code, and a control command code.

[0003] Prior art CN117042253B "An intelligent LED lamp, control system and method", the method includes: an identification module is used to identify the location information of at least one user in the current space through a human body sensor; a scene determination module is used to determine at least one user's current scene based on at least one location information, current space information and at least one of current time information, and the user's current scene includes the actual usage type and / or lighting purpose of the current space; a parameter determination module is used to determine the lighting adjustment parameters based on the current lighting effect and at least one user's current scene.

[0004] As people's quality of life improves, the demand for lighting is no longer limited to basic lighting functions, but a pursuit of more comfortable, energy-saving and intelligent lighting experience. Although existing LED lamps have the advantage of energy saving, there is still a lot of room for improvement in intelligence. Traditional lighting control systems often rely on fixed schedules or simple environmental sensor inputs and cannot achieve a truly personalized lighting experience. Summary of the invention

[0005] In order to solve the above technical problems, the object of the present invention is to provide an adaptive LED lamp control method, comprising the following steps: Step s1: constructing an LED lamp control platform, wherein the LED lamp control platform is communicatively connected with a plurality of indoor LED lamp terminals and a user terminal; Step s2: performing motion feature recognition on the video stream data collected by the LED lamp terminal to obtain the motion information of the target user, and then performing face feature recognition on the video stream data collected by the LED lamp terminal to mark the target user as a registered user or a visitor user; Step s3: When the target user is a tourist user, the control parameters of the LED lamp terminal in the area where the tourist user is located are obtained according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node; Step s4: construct a user command prediction model. When the target user is a registered user, the control parameters of the LED lamp terminal in the area where the registered user is located are obtained based on the multi-source heterogeneous data, action information and user command prediction model of the current control cycle; Step s5: When a registered user sends a comfort score through a user terminal in the current control cycle and does not send command data, the control parameters of the LED lamp terminal in the area where the registered user is located are obtained based on the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node.

[0006] Furthermore, a LED lamp control platform is constructed, and the process of the LED lamp control platform communicating with a plurality of indoor LED lamp terminals and a user terminal includes: An LED lamp control platform is constructed. The LED lamp control platform is communicatively connected to a plurality of blockchain nodes, and the plurality of blockchain nodes are interconnected to form a blockchain network. The blockchain nodes are communicatively connected to the LED lamp terminal and the user terminal to obtain multi-source heterogeneous data collected by the LED lamp terminal and the instruction data, comfort score and face recognition data sent by the user terminal and mark the collection time, and set a control period T. The multi-source heterogeneous data includes light intensity, sound data, video stream data and electricity price information.

[0007] Furthermore, the process of performing motion feature recognition on the video stream data collected by the LED lamp terminal to obtain the motion information of the target user includes: Obtain video stream data collected by the LED lamp terminal, convert the video stream data into n consecutive video segments of duration T, convert each video segment into a corresponding video frame image sequence, use the OpenPose algorithm to perform posture estimation on the video frame image sequence, obtain the human body key point coordinates of the target user in each video frame image, obtain the target user's skeleton data based on the human body key point indicators, and extract feature vectors based on the human body key point coordinates and skeleton data; Construct an action recognition model based on deep learning, obtain preset action data as a training set and a test set, input the training set into the action recognition model for training until the loss function training is stable, and save the model parameters, test the action recognition model with the test set until it meets the preset requirements, output the trained action recognition model, input the feature vector into the action recognition model, and output the action information of the target user.

[0008] Furthermore, the process of performing facial feature recognition on the video stream data collected by the LED lamp terminal and marking the target user as a registered user or a visitor user includes: Fisherfaces is used to build a facial feature recognition model, and the video frame image is input into the facial feature recognition model for feature extraction to obtain the facial feature vector of the target user to be recognized in the video frame image; Build a face database, obtain face recognition data from several blockchain nodes, import it into the face feature recognition model for feature extraction, obtain face feature vectors and store them in the face database; The facial feature vector to be identified is compared with the facial feature vector in the face database for facial feature vector similarity, and the similarity between each facial feature vector and the facial verification vector to be identified is obtained. If there is a facial feature vector and the facial verification vector to be identified whose similarity is greater than or equal to a preset similarity threshold, the user terminal corresponding to the facial feature vector is obtained, and the target user is marked as a registered user according to the user terminal. If the similarity between each facial feature vector and the facial verification vector to be identified is less than the preset similarity threshold, the target user is marked as a tourist user.

[0009] Furthermore, when the target user is a tourist user, the process of obtaining the control parameters of the LED lamp terminal in the area where the tourist user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node includes: A feature matching set of tourist users is constructed based on the light intensity, sound data, electricity price information and action information of tourist users in the current control cycle, and the action information, light intensity, sound data, electricity price information, command data and comfort score in several historical control cycles stored in the blockchain node are obtained. A feature matching set of several historical control cycles is constructed based on the action information, light intensity, sound data and electricity price information in several historical control cycles. The feature matching set of tourist users is feature matched with the feature matching sets of several historical control cycles to obtain the feature matching degree of each historical control cycle and the tourist user, and the feature matching degree of each historical control cycle is compared with a preset feature matching degree threshold, and the historical control cycle with a feature matching degree greater than the feature matching degree threshold is screened out, and the historical control cycle is marked as the first historical control cycle, and the command data and comfort score corresponding to the first historical control cycle are obtained, and the command data corresponding to the first historical control cycle with the highest comfort score is obtained. According to the command data, the control parameters of the LED lamp terminal in the current control cycle are obtained.

[0010] Furthermore, a user command prediction model is constructed. When the target user is a registered user, the process of obtaining the control parameters of the LED lamp terminal in the area where the registered user is located according to the multi-source heterogeneous data, action information and user command prediction model of the current control cycle includes: Obtain the action information, light intensity, sound data, electricity price information and command data of each registered user in several historical control cycles stored in the blockchain node, build a user command prediction model based on the K-Means clustering algorithm and the logistic regression model, use the action information, light intensity, sound data, electricity price information and command data of the target user of each registered user in several historical control cycles as training data, train the user command prediction model, and output the trained user command prediction model; Inputting the action information, light intensity, sound data and electricity price information of the registered users in the current control cycle into the user command prediction model, outputting the predicted command data according to the user command prediction model, and obtaining the control parameters of the LED lamp terminal in the current control cycle according to the predicted command data; When a registered user sends command data through the user terminal in the current control cycle, the command data sent by the user terminal is obtained to obtain the control parameters of the LED lamp terminal in the current control cycle. At the same time, the command data sent by the user terminal in the current control cycle and the action information, light intensity, sound data, and electricity price information of the registered user are used as training data to re-train the user command prediction model.

[0011] Furthermore, when a registered user sends a comfort score through a user terminal in the current control cycle and does not send command data, the process of obtaining the control parameters of the LED lamp terminal in the area where the registered user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node includes: The comfort score sent by the user terminal is compared with a preset comfort score threshold. If the comfort score is less than the comfort score threshold, the action information, light intensity, sound data, and electricity price information of the registered user in the current control cycle are used as a feature matching set, and the action information, light intensity, sound data, and electricity price information of the registered user in several historical control cycles are obtained. According to the action information, light intensity, sound data, and electricity price information in the several historical control cycles, feature matching sets of several historical control cycles are constructed, and the feature matching sets of the registered user are feature matched with the feature matching sets of several historical control cycles to obtain the feature matching degree of each historical control cycle and the registered user, and the feature matching degree of each historical control cycle is compared with a preset feature matching degree threshold, and the historical control cycle with a feature matching degree greater than the feature matching degree threshold is screened out, and the historical control cycle is marked as a second historical control cycle, and the instruction data and the comfort score corresponding to the second historical control cycle are obtained, and the instruction data corresponding to the second historical control cycle with the highest comfort score is obtained, and the control parameters of the LED lamp terminal in the current control cycle are obtained according to the instruction data.

[0012] An adaptive LED lamp control system includes a monitoring center, wherein the monitoring center is communicatively connected to a platform construction module, a data processing module, a first control module, a second control module, and a third control module; The platform building module is used to build an LED lamp control platform; The data processing module is used to perform motion feature recognition on the video stream data collected by the LED lamp terminal, obtain the motion information of the target user, and then perform face feature recognition on the video stream data collected by the LED lamp terminal, and mark the target user as a registered user or a visitor user; The first control module is used to obtain the control parameters of the LED lamp terminal in the area where the tourist user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node when the target user is a tourist user; The second control module is used to build a user command prediction model. When the target user is a registered user, the control parameters of the LED lamp terminal in the area where the registered user is located are obtained based on the multi-source heterogeneous data, action information and user command prediction model of the current control cycle; The third control module is used to obtain the control parameters of the LED lamp terminal in the area where the registered user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node when the registered user sends a comfort score through the user terminal and does not send command data in the current control cycle.

[0013] Compared with the prior art, the present invention has the following beneficial effects: By combining the motion feature recognition of video stream data, face feature recognition, and user command prediction models, etc., the intelligent control of LED lamps is realized. This method not only improves the user's comfort, but also can dynamically adjust the lighting parameters according to the status and preferences of different users, so as to achieve the purpose of energy saving and emission reduction. At the same time, by identifying the identity of the target user (registered user or visitor user), the system can automatically adjust the lighting parameters according to the user's preferences and historical behavior habits. For example, for registered users, the system can adjust the brightness and color temperature according to the user's preferences to provide a personalized lighting experience. The system can automatically adjust the lighting parameters according to the user's behavior pattern and environmental conditions to avoid unnecessary energy consumption. For example, when the user leaves the room, the system can automatically reduce the brightness or turn off the lights to save energy. By real-time monitoring of the user's motion information and environmental conditions, the system can dynamically adjust the lighting parameters to ensure the user's comfort in different scenarios. For example, when the user is reading or working, the system can increase the brightness and adjust the color temperature to provide a better visual experience. At the same time, the security and integrity of the data are guaranteed by storing multi-source heterogeneous data through blockchain technology. The distributed nature of the blockchain also improves the reliability of the system. Even if a node has a problem, it will not affect the normal operation of the overall system. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a schematic diagram of an adaptive LED lamp control method according to an embodiment of the present application.

[0015] Figure 2 This is a schematic diagram of an adaptive LED lamp control system according to an embodiment of the present application. DETAILED DESCRIPTION

[0016] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0017] like Figure 1 As shown, an adaptive LED lamp control method includes the following steps: Step s1: constructing an LED lamp control platform, wherein the LED lamp control platform is communicatively connected with a plurality of indoor LED lamp terminals and a user terminal; Step s2: performing motion feature recognition on the video stream data collected by the LED lamp terminal to obtain the motion information of the target user, and then performing face feature recognition on the video stream data collected by the LED lamp terminal to mark the target user as a registered user or a visitor user; Step s3: When the target user is a tourist user, the control parameters of the LED lamp terminal in the area where the tourist user is located are obtained according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node; Step s4: construct a user command prediction model. When the target user is a registered user, the control parameters of the LED lamp terminal in the area where the registered user is located are obtained based on the multi-source heterogeneous data, action information and user command prediction model of the current control cycle; Step s5: When a registered user sends a comfort score through a user terminal in the current control cycle and does not send command data, the control parameters of the LED lamp terminal in the area where the registered user is located are obtained based on the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node.

[0018] It should be further explained that, in the specific implementation process, the LED lamp control platform is constructed, and the process of the LED lamp control platform communicating with a plurality of indoor LED lamp terminals and user terminals includes: An LED lamp control platform is constructed. The LED lamp control platform is communicatively connected to a plurality of blockchain nodes, and the plurality of blockchain nodes are interconnected to form a blockchain network. The blockchain nodes are communicatively connected to LED lamp terminals and user terminals to obtain multi-source heterogeneous data collected by LED lamp terminals and instruction data, comfort scores, and face recognition data sent by user terminals (including smart phones, tablet computers, and smart watches, etc.), mark the collection time, and set a control period T. The multi-source heterogeneous data includes light intensity (detecting the natural light intensity around the LED lamp terminal through the ambient light sensor of the LED lamp terminal), sound data (detecting environmental noise through a microphone or an acoustic wave sensor), video stream data (using a camera to detect video data in the room), and electricity price information.

[0019] It should be further explained that, in the specific implementation process, the specific process of the blockchain node acquiring the multi-source heterogeneous data collected by the LED lamp terminal, the command data sent by the user terminal, and the face recognition data is as follows: The public key and private key of each blockchain node are preset through an asymmetric encryption algorithm, and the multi-source heterogeneous data collected by the LED lamp terminal or the instruction data and face recognition data sent by the user terminal are encrypted by the public key of the blockchain node connected to the LED lamp terminal or the user terminal, and then transmitted to the blockchain node. The blockchain node decrypts the received encrypted multi-source heterogeneous data or instruction data and face recognition data through the private key. When the decryption is successful, the multi-source heterogeneous data or instruction data and face recognition data are packaged into a new block, and the new block is broadcast to the entire blockchain network. A consensus mechanism is preset, and other blockchain nodes in the blockchain network verify the block data of the new block according to the consensus mechanism. After the verification of the new block is passed, the new block is added to the end of the blockchain, and the block information in the new block is updated to the local blockchain copies of all blockchain nodes.

[0020] It should be further explained that, in the specific implementation process, the process of performing motion feature recognition on the video stream data collected by the LED lamp terminal and obtaining the motion information of the target user includes: The video stream data collected by the LED lamp terminal is obtained, and the video stream data is converted into n continuous video segments of a duration of T, and each video segment is converted into a corresponding video frame image sequence. The OpenPose algorithm is used to estimate the posture of the video frame image sequence, and the human body key point coordinates of the target user in each video frame image are obtained. The skeleton data of the target user is obtained according to the human body key point index, and the feature vector is extracted according to the human body key point coordinates and the skeleton data. The OpenPose algorithm preprocesses the image to be processed, such as scaling, cropping, etc., to meet the requirements of subsequent processing, and uses the skeleton point positioning network to extract features from the preprocessed image and generate a confidence map for each skeleton point. Each pixel value in the confidence map represents the probability that the position is a certain skeleton point. Subsequently, the affinity detection network is used to analyze the local area around the skeleton point to generate an affinity map for determining which skeleton points should be connected to form a complete human skeleton. Finally, the confidence map generated by the skeleton point positioning network and the affinity map generated by the affinity detection network are summarized, and the final skeleton point position is determined by a certain algorithm (such as non-maximum suppression, greedy algorithm, etc.), and these points are connected to form a complete skeleton. Construct an action recognition model based on deep learning, obtain preset action data as a training set and a test set, input the training set into the action recognition model for training until the loss function training is stable, and save the model parameters, test the action recognition model with the test set until it meets the preset requirements, output the trained action recognition model, input the feature vector into the action recognition model, and output the action information of the target user.

[0021] It should be further explained that, in the specific implementation process, the process of performing facial feature recognition on the video stream data collected by the LED lamp terminal and marking the target user as a registered user or a visitor user includes: Fisherfaces is used to build a facial feature recognition model, and the video frame image is input into the facial feature recognition model for feature extraction to obtain the facial feature vector of the target user to be recognized in the video frame image; Build a face database, obtain face recognition data from several blockchain nodes, import it into the face feature recognition model for feature extraction, obtain face feature vectors and store them in the face database; The facial feature vector to be identified is compared with the facial feature vector in the face database for facial feature vector similarity, and the similarity between each facial feature vector and the facial verification vector to be identified is obtained. If there is a facial feature vector and the facial verification vector to be identified whose similarity is greater than or equal to a preset similarity threshold, the user terminal corresponding to the facial feature vector is obtained, and the target user is marked as a registered user according to the user terminal. If the similarity between each facial feature vector and the facial verification vector to be identified is less than the preset similarity threshold, the target user is marked as a tourist user.

[0022] It should be further explained that, in the specific implementation process, when the target user is a tourist user, the process of obtaining the control parameters of the LED lamp terminal in the area where the tourist user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node includes: A feature matching set of tourist users is constructed based on the light intensity, sound data, electricity price information and action information of tourist users in the current control cycle, and the action information, light intensity, sound data, electricity price information, command data and comfort score in several historical control cycles stored in the blockchain node are obtained. A feature matching set of several historical control cycles is constructed based on the action information, light intensity, sound data and electricity price information in several historical control cycles. The feature matching set of tourist users is feature matched with the feature matching sets of several historical control cycles to obtain the feature matching degree of each historical control cycle and the tourist user, and the feature matching degree of each historical control cycle is compared with a preset feature matching degree threshold, and the historical control cycle with a feature matching degree greater than the feature matching degree threshold is screened out, and the historical control cycle is marked as the first historical control cycle, and the command data and comfort score corresponding to the first historical control cycle are obtained, and the command data corresponding to the first historical control cycle with the highest comfort score is obtained. According to the command data, the control parameters of the LED lamp terminal in the current control cycle are obtained.

[0023] It should be further explained that, in the specific implementation process, a user command prediction model is constructed. When the target user is a registered user, the process of obtaining the control parameters of the LED lamp terminal in the area where the registered user is located according to the multi-source heterogeneous data, action information and user command prediction model of the current control cycle includes: Obtain the action information, light intensity, sound data, electricity price information and command data of each registered user in several historical control cycles stored in the blockchain node, build a user command prediction model based on the K-Means clustering algorithm and the logistic regression model, use the action information, light intensity, sound data, electricity price information and command data of the target user of each registered user in several historical control cycles as training data, train the user command prediction model, and output the trained user command prediction model; Inputting the action information, light intensity, sound data and electricity price information of the registered users in the current control cycle into the user command prediction model, outputting the predicted command data according to the user command prediction model, and obtaining the control parameters of the LED lamp terminal in the current control cycle according to the predicted command data; When a registered user sends command data through the user terminal in the current control cycle, the command data sent by the user terminal is obtained to obtain the control parameters of the LED lamp terminal in the current control cycle. At the same time, the command data sent by the user terminal in the current control cycle and the action information, light intensity, sound data, and electricity price information of the registered user are used as training data to re-train the user command prediction model.

[0024] A cluster analysis is performed on the target user's action information, light intensity, sound data, electricity price information, and command data. The purpose is to classify similar command data into one category. Based on the clustering results, a logistic regression model is established to predict the probability of a user issuing a specific command under given conditions. The above model is trained using historical data, and the accuracy and generalization ability of the model are evaluated through cross-validation and other methods. The trained model is applied to new data sets to predict in real time the probability of a command that a user may issue under current conditions, and the control strategy of the LED lamps is optimized accordingly.

[0025] It should be further explained that, in the specific implementation process, when a registered user sends a comfort score through a user terminal in the current control cycle and does not send command data, the process of obtaining the control parameters of the LED lamp terminal in the area where the registered user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node includes: The comfort score sent by the user terminal is compared with a preset comfort score threshold. If the comfort score is less than the comfort score threshold, the action information, light intensity, sound data, and electricity price information of the registered user in the current control cycle are used as a feature matching set, and the action information, light intensity, sound data, and electricity price information of the registered user in several historical control cycles are obtained. According to the action information, light intensity, sound data, and electricity price information in the several historical control cycles, feature matching sets of several historical control cycles are constructed, and the feature matching sets of the registered user are feature matched with the feature matching sets of several historical control cycles to obtain the feature matching degree of each historical control cycle and the registered user, and the feature matching degree of each historical control cycle is compared with a preset feature matching degree threshold, and the historical control cycle with a feature matching degree greater than the feature matching degree threshold is screened out, and the historical control cycle is marked as a second historical control cycle, and the instruction data and the comfort score corresponding to the second historical control cycle are obtained, and the instruction data corresponding to the second historical control cycle with the highest comfort score is obtained, and the control parameters of the LED lamp terminal in the current control cycle are obtained according to the instruction data.

[0026] like Figure 2 As shown, an adaptive LED lamp control system includes a monitoring center, wherein the monitoring center is communicatively connected with a platform construction module, a data processing module, a first control module, a second control module and a third control module; The platform building module is used to build an LED lamp control platform; The data processing module is used to perform motion feature recognition on the video stream data collected by the LED lamp terminal, obtain the motion information of the target user, and then perform face feature recognition on the video stream data collected by the LED lamp terminal, and mark the target user as a registered user or a visitor user; The first control module is used to obtain the control parameters of the LED lamp terminal in the area where the tourist user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node when the target user is a tourist user; The second control module is used to build a user command prediction model. When the target user is a registered user, the control parameters of the LED lamp terminal in the area where the registered user is located are obtained based on the multi-source heterogeneous data, action information and user command prediction model of the current control cycle; The third control module is used to obtain the control parameters of the LED lamp terminal in the area where the registered user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node when the registered user sends a comfort score through the user terminal and does not send command data in the current control cycle.

[0027] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An adaptive LED lamp control method, characterized in that: The following steps are involved: Step s1: constructing an LED lamp control platform, wherein the LED lamp control platform is communicatively connected with a plurality of indoor LED lamp terminals and a user terminal; Step s2: performing motion feature recognition on the video stream data collected by the LED lamp terminal to obtain the motion information of the target user, and then performing face feature recognition on the video stream data collected by the LED lamp terminal to mark the target user as a registered user or a visitor user; Step s3: When the target user is a tourist user, the control parameters of the LED lamp terminal in the area where the tourist user is located are obtained according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node; Step s4: construct a user command prediction model. When the target user is a registered user, the control parameters of the LED lamp terminal in the area where the registered user is located are obtained based on the multi-source heterogeneous data, action information and user command prediction model of the current control cycle; Step s5: When a registered user sends a comfort score through a user terminal in the current control cycle and does not send command data, the control parameters of the LED lamp terminal in the area where the registered user is located are obtained based on the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node.

2. The adaptive LED lamp control method according to claim 1, characterized in that: The process of constructing an LED lamp control platform, wherein the LED lamp control platform communicates with a plurality of indoor LED lamp terminals and a user terminal includes: An LED lamp control platform is constructed. The LED lamp control platform is communicatively connected to a plurality of blockchain nodes, and the plurality of blockchain nodes are interconnected to form a blockchain network. The blockchain nodes are communicatively connected to the LED lamp terminal and the user terminal to obtain multi-source heterogeneous data collected by the LED lamp terminal and the instruction data, comfort score and face recognition data sent by the user terminal and mark the collection time, and set a control period T. The multi-source heterogeneous data includes light intensity, sound data, video stream data and electricity price information.

3. The adaptive LED lamp control method according to claim 2, characterized in that: The process of performing motion feature recognition on the video stream data collected by the LED lamp terminal and obtaining the motion information of the target user includes: Obtain video stream data collected by the LED lamp terminal, convert the video stream data into n consecutive video segments of duration T, convert each video segment into a corresponding video frame image sequence, use the OpenPose algorithm to perform posture estimation on the video frame image sequence, obtain the human body key point coordinates of the target user in each video frame image, obtain the target user's skeleton data based on the human body key point indicators, and extract feature vectors based on the human body key point coordinates and skeleton data; Construct an action recognition model based on deep learning, obtain preset action data as a training set and a test set, input the training set into the action recognition model for training until the loss function training is stable, and save the model parameters, test the action recognition model with the test set until it meets the preset requirements, output the trained action recognition model, input the feature vector into the action recognition model, and output the action information of the target user.

4. The adaptive LED lamp control method according to claim 3, characterized in that: The process of performing facial feature recognition on the video stream data collected by the LED lamp terminal and marking the target user as a registered user or a visitor user includes: Fisherfaces is used to build a facial feature recognition model, and the video frame image is input into the facial feature recognition model for feature extraction to obtain the facial feature vector of the target user to be recognized in the video frame image; Build a face database, obtain face recognition data from several blockchain nodes, import it into the face feature recognition model for feature extraction, obtain face feature vectors and store them in the face database; The facial feature vector to be identified is compared with the facial feature vector in the face database for facial feature vector similarity, and the similarity between each facial feature vector and the facial verification vector to be identified is obtained. If there is a facial feature vector and the facial verification vector to be identified whose similarity is greater than or equal to a preset similarity threshold, the user terminal corresponding to the facial feature vector is obtained, and the target user is marked as a registered user according to the user terminal. If the similarity between each facial feature vector and the facial verification vector to be identified is less than the preset similarity threshold, the target user is marked as a tourist user.

5. The adaptive LED lamp control method according to claim 4, characterized in that: When the target user is a tourist user, the process of obtaining the control parameters of the LED lamp terminal in the area where the tourist user is located, based on the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node, includes: A feature matching set of tourist users is constructed based on the light intensity, sound data, electricity price information and action information of tourist users in the current control cycle, and the action information, light intensity, sound data, electricity price information, command data and comfort score in several historical control cycles stored in the blockchain node are obtained. A feature matching set of several historical control cycles is constructed based on the action information, light intensity, sound data and electricity price information in several historical control cycles. The feature matching set of tourist users is feature matched with the feature matching sets of several historical control cycles to obtain the feature matching degree of each historical control cycle and the tourist user, and the feature matching degree of each historical control cycle is compared with a preset feature matching degree threshold, and the historical control cycle with a feature matching degree greater than the feature matching degree threshold is screened out, and the historical control cycle is marked as the first historical control cycle, and the command data and comfort score corresponding to the first historical control cycle are obtained, and the command data corresponding to the first historical control cycle with the highest comfort score is obtained. According to the command data, the control parameters of the LED lamp terminal in the current control cycle are obtained.

6. The adaptive LED lamp control method according to claim 5, characterized in that: Construct a user command prediction model. When the target user is a registered user, the process of obtaining the control parameters of the LED lamp terminal in the area where the registered user is located based on the multi-source heterogeneous data, action information and user command prediction model of the current control cycle includes: Obtain the action information, light intensity, sound data, electricity price information and command data of each registered user in several historical control cycles stored in the blockchain node, build a user command prediction model based on the K-Means clustering algorithm and the logistic regression model, use the action information, light intensity, sound data, electricity price information and command data of the target user of each registered user in several historical control cycles as training data, train the user command prediction model, and output the trained user command prediction model; Inputting the action information, light intensity, sound data and electricity price information of the registered users in the current control cycle into the user command prediction model, outputting the predicted command data according to the user command prediction model, and obtaining the control parameters of the LED lamp terminal in the current control cycle according to the predicted command data; When a registered user sends command data through the user terminal in the current control cycle, the command data sent by the user terminal is obtained to obtain the control parameters of the LED lamp terminal in the current control cycle. At the same time, the command data sent by the user terminal in the current control cycle and the action information, light intensity, sound data, and electricity price information of the registered user are used as training data to re-train the user command prediction model.

7. The adaptive LED lamp control method according to claim 6, characterized in that: When a registered user sends a comfort score through a user terminal in the current control cycle and does not send command data, the process of obtaining the control parameters of the LED lamp terminal in the area where the registered user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node includes: The comfort score sent by the user terminal is compared with a preset comfort score threshold. If the comfort score is less than the comfort score threshold, the action information, light intensity, sound data, and electricity price information of the registered user in the current control cycle are used as a feature matching set, and the action information, light intensity, sound data, and electricity price information of the registered user in several historical control cycles are obtained. According to the action information, light intensity, sound data, and electricity price information in the several historical control cycles, feature matching sets of several historical control cycles are constructed, and the feature matching sets of the registered user are feature matched with the feature matching sets of several historical control cycles to obtain the feature matching degree of each historical control cycle and the registered user, and the feature matching degree of each historical control cycle is compared with a preset feature matching degree threshold, and the historical control cycle with a feature matching degree greater than the feature matching degree threshold is screened out, and the historical control cycle is marked as a second historical control cycle, and the instruction data and the comfort score corresponding to the second historical control cycle are obtained, and the instruction data corresponding to the second historical control cycle with the highest comfort score is obtained, and the control parameters of the LED lamp terminal in the current control cycle are obtained according to the instruction data.

8. An adaptive LED lamp control system, specifically applied to an adaptive LED lamp control method according to any one of claims 1 to 7, characterized in that: It includes a monitoring center, wherein the monitoring center is communicatively connected with a platform construction module, a data processing module, a first control module, a second control module and a third control module; The platform building module is used to build an LED lamp control platform; The data processing module is used to perform motion feature recognition on the video stream data collected by the LED lamp terminal, obtain the motion information of the target user, and then perform face feature recognition on the video stream data collected by the LED lamp terminal, and mark the target user as a registered user or a visitor user; The first control module is used to obtain the control parameters of the LED lamp terminal in the area where the tourist user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node when the target user is a tourist user; The second control module is used to build a user command prediction model. When the target user is a registered user, the control parameters of the LED lamp terminal in the area where the registered user is located are obtained based on the multi-source heterogeneous data, action information and user command prediction model of the current control cycle; The third control module is used to obtain the control parameters of the LED lamp terminal in the area where the registered user is located according to the multi-source heterogeneous data and action information stored in the current control cycle and the blockchain node when the registered user sends a comfort score through the user terminal and does not send command data in the current control cycle.

Images